A three-dimensional radiative transfer model to interpret ranging scatterometer measurements from a pine forest

Abstract

This paper addresses the propagation of microwaves in forested media. It interprets the vertical distribution of backscatter acquired by a ranging scatterometer in a pine forest at X band (9.5 GHz) and vertical incidence by means of modelling. A plane parallel radiative transfer model for the forest canopy exhibits a large discrepancy between the measured and estimated distribution. This is attributed to failure to represent the horizontal heterogeneity in the forest. Taking into account the height variability of the trees significantly improves the estimated distribution at the top of the canopy but the backscatter from the lower layers and the ground is still underestimated. To deal with this, a fully three-dimensional (3D) radiative transfer model is developed. By modelling precisely the propagation of the wave through the medium, the 3D approach allows accurate estimation of the ground backscatter and slightly improves the estimated backscatter from the lower layers. It is demonstrated that 3D modelling is required to estimate accurately the vertical distribution of backscatter, total backscatter and the height of the centre of backscatter in the experimental data. These results have implications for the interpretation of more general measurements scenarios, especially as regards forest height measurement by radar interferometry.